It is pointed out the deficiency in Haken's paper in which he discussed the change of entropy in the process of nonequilibrium phase transition, and an improved suggestion is made.